Objectives  Multi-scale boundary information can be obtained using potential field data at different scales, which having an important practical value on account of the tectonic division, geological mapping orebody delineation and the research of crystalline basement.

  Methods  We use the bi-dimensional empirical mode decomposition (BEMD) and the wavelet modulus maximum (WMM) methods to obtain the multi-scale boundary detection of potential field data.Firstly, it is shown that the BEMD method can realize the multi-scale decomposition of potential field data, and the anomalies at different scales possess practical geological significance. Secondly, the model test established by this paper shows that WMM method has the ability of resisting the noise interference so as to detect the boundary of the combined model about the distribution, location and direction clearly and accurately using the gravity anomalies. Finally, this paper also achieves the results of multi-scale decomposition of the Bouguer gravity anomalies in the Three Gorges region by BEMD algorithm, and estimates the approximate depth of the buried source by the radial logarithmic power spectrum of each BIMF (bi-dimensional intrinsic mode function)and residue, and we deal with each BIMF using WMM method so as to obtain the distribution characteristic of the structure and rock mass in different depths.

  Results and Conclusions  The results of boundary detection at all levels are analyzed and explained according to the regional geology structure feature, the advantage and effectiveness of this method is verified in the end.